Basically, before Phylip, Hennig86, and PAUP (no "*", just "PAUP"), cladistic analyses had to be done by hand. For analyses with more than 6 or 7 operational taxonomic units and a dozen or so characters, this would be very "computationally" (i.e., pen & paper) time consuming.
Hennig, who had plenty of time, did a few such analyses.After him and well into the 1990s, however, many people misused the term, came up with a phylogenetic hypothesis, drew a branching diagram (which they mistakenly called "cladogram") to illustrate it, and showed where on the tree, according to their hypothesis, a few characters they were interested in changed states -- often missing where else in the tree those changes happened, often defining only one of the states, and so on. That's a fairly explicit way of presenting a phylogenetic hypothesis, definitely much better than what was done before, but still not a way to _test_ phylogenetic hypotheses with any precision.